AI News · $100M+ rounds ·
TypeSafe AI raises $870 million at $7.5 billion valuation for non-text model Jev
TypeSafe AI, the developer of the Jev AI model, raised $870 million at a $7.5 billion valuation, TechCrunch reported. Andreessen Horowitz led the round, with Sequoia and existing investor DCVC participating. Jev launched on September 15, 2026 and outputs probabilities rather than text. The company says a third of Fortune 500 companies already use it.
Key points
- TypeSafe AI raised $870 million at a $7.5 billion valuation.
- Andreessen Horowitz led; Sequoia and existing investor DCVC participated.
- Jev launched September 15, 2026 and outputs probabilities instead of text.
- The company says a third of Fortune 500 companies already use Jev.
- Founders are former OpenAI researcher Diogo Almeida, Sasha Sheng and Erik Gafni.
What happened: TypeSafe AI, the company behind the Jev AI model, raised $870 million at a $7.5 billion valuation, TechCrunch reported on October 9. Andreessen Horowitz led the round. Sequoia participated, and existing investor DCVC also took part. The raise comes only weeks after Jev was released on September 15, 2026.
The details: Jev uses a transformer architecture, the same basic design behind most modern AI models, but it is not a large language model. Instead of producing text, it outputs probabilities, which the company calls "calibrated decisions." TypeSafe says Jev is "significantly faster and uses far fewer tokens than LLMs," and it positions the model for automating tasks rather than generating text or code. The company says a third of Fortune 500 companies already use Jev.
TypeSafe was co-founded in 2024 by Diogo Almeida, a former OpenAI researcher, Sasha Sheng, a former Meta research engineer, and Erik Gafni, an engineer and entrepreneur. Speaking to TechCrunch the previous month, Almeida said, "We have been super good at human language for four years," and argued that for automation, "computers speak a different language."
Background: Most business AI tools today are built on large language models that read and write text. Using them to make routine decisions, such as classifying a record or choosing the next step in a workflow, can be slow and costly because every step consumes tokens, the units of text that providers charge for. A model that returns a direct probability for a decision aims to cut that cost and delay. TechCrunch's report did not include independent benchmarks of Jev's speed or accuracy, and the company has not disclosed pricing details in the report.
Who it affects: Operations, finance and data teams that run high-volume automated decisions are the most obvious audience. Buyers should test whether calibrated probabilities from Jev fit their existing systems and how they compare with current language model setups on accuracy and cost.
What to watch: With this much new capital, watch for how TypeSafe expands access, what pricing it publishes, and whether outside evaluations confirm its speed and token claims. The size of the round shows investors are willing to back AI architectures that move away from text generation for business automation.
Our take
Jev targets task automation and decisions rather than writing, so teams automating structured workflows may get a faster, cheaper option than routing every step through a large language model.